Most Cost-Effective EC2 Scaling for Sudden Seasonal Sale Traffic?

Why the Answer Is Correct

Option D directly satisfies both halves of the requirement: the website must absorb sudden traffic increases, and it must do so MOST cost-effectively. An Auto Scaling group that scales out when traffic rises and scales in when the sale ends means the company pays only for the EC2 capacity actually consumed, plus it keeps the multi-AZ architecture intact. The launch template is not decoration — it captures the preconfigured AMI, instance type, security groups, and user data so new instances launch ready to serve traffic in minutes without manual setup, which is what makes rapid scale-out viable during a flash sale. As JA2018 put it, using a launch template with a preconfigured AMI "ensures new instances are quickly provisioned". Nothing in the scenario suggests predictable, steady-state demand that would justify pre-buying capacity, so elastic scaling on demand is the cheapest correct design.

Why the Other Options Are Wrong

Option A is an anti-pattern: a stopped EC2 instance still counts toward the Auto Scaling group's desired capacity, and the group treats an impaired instance as unhealthy and replaces it rather than "using" it to scale out — so the design does not work as described. Option B sizes the minimum capacity to handle peak traffic without scaling out, which means paying for peak-capable infrastructure 24/7 for a sale that lasts days; that is the opposite of cost-effective. Option C is technically plausible but not the cheapest: Amazon ElastiCache nodes bill continuously, dynamic content from an ecommerce site often has low cache hit ratios, and the answer even requires running enough instances "to populate CloudFront and ElastiCache" before scaling in, which front-loads cost the scenario does not need. As buzzinmumbai summarized, "C is not cost effective to use elasticache".

Community Comment Notes

Consensus is overwhelmingly for D — the recommended answer received all recorded votes from commenters such as Andy_09, stephensimudemy, and geraltRebo. sandordini made the key judgment call plainly: CloudFront "could be a good idea, but it seems to be a simple scaling scenario". buzzinmumbai added the cost argument against caching. One commenter (viejito) argued in Spanish for C on the grounds that caching reduces the load on EC2, which is true for performance but ignores the ongoing ElastiCache cost for a short seasonal spike, and geraltRebo posted both C and D, showing the caching distractor is attractive but not the cost-optimal reading.